w/youtube-dl
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

[thvideo] Support mobile URLs as well

Browse source
This commit is contained in:
Philipp Hagemeister 2014-09-21 16:32:38 +02:00
parent 3ef7d11acd
commit df8f53f752
1 changed files with 12 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

20
youtube_dl/extractor/thvideo.py
View file

@ -10,7 +10,7 @@ from ..utils import (
class THVideoIE(InfoExtractor): class THVideoIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?thvideo\.tv/v/th(?P<id>[0-9]+)' _VALID_URL = r'http://(?:www\.)?thvideo\.tv/(?:v/th|mobile\.php\?cid=)(?P<id>[0-9]+)'
_TEST = { _TEST = {
'url': 'http://thvideo.tv/v/th1987/', 'url': 'http://thvideo.tv/v/th1987/',
'md5': 'fa107b1f73817e325e9433505a70db50', 'md5': 'fa107b1f73817e325e9433505a70db50',
@ -30,18 +30,22 @@ class THVideoIE(InfoExtractor):
video_id = mobj.group('id') video_id = mobj.group('id')
# extract download link from mobile player page # extract download link from mobile player page
webpage_player = self._download_webpage('http://thvideo.tv/mobile.php?cid=%s-0' % video_id, video_id) webpage_player = self._download_webpage(
video_url = self._html_search_regex(r'<source src="(.*?)" type', webpage_player, 'video url') 'http://thvideo.tv/mobile.php?cid=%s-0' % (video_id),
video_id, note='Downloading video source page')
video_url = self._html_search_regex(
r'<source src="(.*?)" type', webpage_player, 'video url')
# extract video info from main page # extract video info from main page
webpage = self._download_webpage(url, video_id) webpage = self._download_webpage(
'http://thvideo.tv/v/th%s' % (video_id), video_id)
title = self._og_search_title(webpage) title = self._og_search_title(webpage)
display_id = 'th%s' % video_id display_id = 'th%s' % video_id
thumbnail = self._og_search_thumbnail(webpage) thumbnail = self._og_search_thumbnail(webpage)
description = self._og_search_description(webpage) description = self._og_search_description(webpage)
upload_date_raw = self._html_search_regex(r'span itemprop="datePublished" content="(.*?)">', webpage, upload_date = unified_strdate(self._html_search_regex(
'upload date', fatal=False) r'span itemprop="datePublished" content="(.*?)">', webpage,
upload_date = unified_strdate(upload_date_raw) 'upload date', fatal=False))
return { return {
'id': video_id, 'id': video_id,
@ -52,4 +56,4 @@ class THVideoIE(InfoExtractor):
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'description': description, 'description': description,
'upload_date': upload_date 'upload_date': upload_date
} }